Agent-based task assignment in multi-hop wireless sensor and robot networks improved by iMesh and auctions

Abstract

In this article, an improvement to the existing agent-based auction aggregation protocol for the task assignment (k-aSAAP) in multi-hop wireless sensor and robot networks is proposed by applying the localized distance-sensitive information mesh based service discovery algorithm (iMesh). The agents are introduced to reduce the communication overhead and further improved by introducing the iMesh service discovery in the following way: the area monitored by sensors is divided into sectors where one static agent, having all information about the robots in that sector, is located in the center. Event information is detected by one of the sensors. That sensor then conducts the lookup process in the sensor matrix to obtain the information about nearby service providers (robots) from the iMesh structure. After the nearest service provider is found, it now forwards the information to the agent in that sector which triggers the auction searching for the robot with the lowest cost among the neighboring agents. The existing agent-based solutions and the iMesh were compared and later combined to improve the percentage of optimal assignments, the average optimal-to-closest distance ratio and to maintain low the number of exchanged messages. The performance of the protocols are shown by the simulation results which show that introducing the iMesh can improve the percentage of optimal assignments from 50 to over 90% for low values of k (number of hops), while keeping the number of average exchanged messages to around 1 per robot. © 2012 IEEE.